Gravitational Wave Observations with Interferometers: Results and Prospects

1
Gravitational Wave Observations with
Interferometers Results and Prospects

• Stan Whitcomb
• for the LIGO Scientific Collaboration
• 2nd Gravitational Wave Phenomenology Workshop
• Penn State University
• 6 November 2003

2
GW interferometers

• TAMA
• First observations Sept 1999
• gt80 duty cycle over DT8/S2 run (Feb-Apr 2002)
• GEO600
• Advanced featuresfused silica suspensions,
  signal recycling, etc.
• Stable lock in signal recycled mode preparing
  for first observations in this mode
• Virgo
• First arm locked Oct 2003
• LIGO
• First full interferometer lock Oct 2000
• Total of three interferometers at two sites--L1,
  H1, H2 (2km)
• Essentially identical orientation

3

Detector Sensitivity Progression

• Steady improvement in LIGO interferometers
• Example Livingston interferometer (L1)

S1
S2

• TAMA interferometer
• Installation of improved seismic isolation,
  planned for next year, should aid at low
  frequencies

4
First LIGO Science Run (S1)

• August 23 - September 9 (400 hours duration)
• 1st coincidence interferometerobservations since
  1989 (100 hour run)
• Three LIGO interferometers, plus GEO (Europe)
  and TAMA (Japan)
• Hardware reliability good for this stage in the
  commissioning
• Longest locked segment for LIGO interferometer
  21 hrs

LLO-4K LHO-4K LHO-2K 3x Coinc. GEO600
Duty cycle 42 58 73 24 97
5
Astrophysical Searches with Interferometer Data

• Compact binary inspiral chirps
• NS-NS waveforms are well described
• BH-BH need better waveforms
• search technique matched templates
• Supernovae / GRBs bursts
• burst signals in coincidence with signals in
  electromagnetic radiation
• prompt alarm ( one hour) with neutrino detectors
• Pulsars in our galaxy periodic
• search for observed neutron stars (frequency,
  doppler shift)
• all sky search (computing challenge)
• r-modes, LMXBs
• Cosmological Signals stochastic
  background

6
Compact Binary Coalescence

• Search technique matched templates
• Neutron Star Neutron Star
• waveforms known with confidence
• Black Hole Black Hole
• need better waveforms
• TAMA DT6 search
• m1m2 lt 10 M?
• LIGO S1 Search
• Discrete set of templates labeled by (m1, m2)
• 1.0 M? lt m1, m2 lt 3.0 M?
• 2110 templates
• At most 3 loss in SNR

7
TAMA Range for Binary Inspirals
(total)
8
Results of S1 Inspiral Search

• Monte Carlo simulation to determine
  efficiency for detecting galactic events
• Simulated Galactic Population includes Milky
  Way, LMC and SMC
• LMC and SMC contribute 12

TAMA
LIGO S1 Upper Limit R lt 170 / yr / MWEG (Milky
Way Equivalent Galaxy)
TAMA DT6 Upper Limit R lt 120 / yr

• Theoretical prediction R 10-4 - 10-6 / yr /
  MWEG (??)
• Potential for improvement
• 100-300 x increase in range (20 Mpc)
• 100 x observation time

9
Short GW Burst Sources

• Known sources -- Supernovas GRBs
• Coincidence with observed electromagnetic
  observations.
• No close events occurred during S1
• Second science run We are analyzing data near
  the very bright and close GRB030329 (both Hanford
  detectors and TAMA operating)

• Unknown phenomena
• Emission of short transients of gravitational
  radiation of unknown waveform (e.g. black hole
  mergers).

10
Unmodelled Burst Search
search for waveforms from sources for which we
cannot currently make an accurate prediction of
the waveform shape.
GOAL
METHODS
Raw Data
Temporal coincidence
Time-domain high pass filter
Time-Frequency Plane
TFCLUSTERS
Cross correlation
Amplitude test
Df
Dt
time
Final event list
11
Determination of Efficiency
Detection efficiency vs. amplitude,
averaged over source direction and
polarization
To determine sensitivity, inject
representative waveforms into actual data and
run through the analysis pipeline
1ms Gaussian burst
12
LIGO S1 Upper Limit ex 1ms gaussian bursts
Excluded regions in rate-amplitude plane

• Not as good as the best bar results to date, due
  to their
• Longer observation time
• Higher sensitivity near 1 kHz
• Broaden parameter space of waveforms searched
• Longer duration bursts
• Astrophysically motivated
• Prospects for improvement
• 300-1000x detector sensitivity
• 300x in observation time
• 3x analysis improvements (?)
• ?x improved gaussianity

90 confidence
13
CW Sources and Search Methods

• Neutron stars in our galaxy
• Search for observed neutron stars (known location
  and frequency)
• Low mass X-ray binaries (known location, rough
  frequency range)
• Unobserved NSs (unknown location, unknown
  frequency)
• Search Challenges
• Frequency modulation of signal due to Earths
  motion
• Amplitude modulation due to the detectors
  antenna pattern
• All sky search represents significant
  computational challenge
• Search methods
• Time Domain
• Computationally easy but best suited to known
  sources
• Frequency Domain
• Best suited for large parameter space searches

14
TAMA Search for SN1987A

• Evidence of modulated emission at 467.5 Hz
• GW emission expected at 935 Hz
• Highest sensitivity region of TAMA300
• DT6 1000 hours of observation in 2001
• Search over 0.1 Hz bandwidth
• Upper limit
• h lt 5 x 10-23
• 99 confidence level

15
Directed Search in LIGO S1
NO DETECTION EXPECTED at S1 sensitivities

• Compare searches using time and frequency domain
  algorithms
• Confront challenge of coherent analysis of
  detectors with different orientations on
  different continents

PSR J19392134 1283.86 Hz
16
S1 Result PSR J19392134

• Upper limit for targeted pulsar
• Comparison of frequency domain and time domain
  searches
• 95 upper limits on h
• Spindown estimate h lt 1.8 x 10-27
• Prospects for improvement
• 100-1000x from detector sensitivity (depending on
  frequency)
• 10x from observation time

IFO Frequentist FDS Bayesian
TDS GEO 1.9 x 10-21
2.2 x 10-21 LLO 2.7 x 10-22
1.4 x 10-22 LHO-2K 4.0 x 10-22
2.4 x 10-22 LHO-4K
5.4 x 10-22 3.3 x 10-22
17
Stochastic Background

• Strength specified by ratio of GW energy density
  to closure density
• Detect by cross-correlating output of two
  interferometer detectors
• Use widely separated detectors to minimize
  correlated environmental noise
• Prospects for improvement in W
• 106 x from detector sensitivity improvements (W
  h2 )
• 10 x from observation time

Hanford - Livingston
LHO 2km-LLO 4km 61 Hours of S1 data WGW (40Hz -
314 Hz) lt 23
18
Stochastic Background measurements and
predictions
results
projected
LIGO
Adv LIGO
19
Second LIGO Science Run (S2)TAMA Data-taking 8
(DT8)

• February 14 April 14, 2002 ( 1400 hours)
• Three LIGO interferometers and TAMA (Japan)
• 10x sensitivity improvement over S1
• Duty cycle similar to S1
• Increased sensitivity did not degrade operation
• Longest locked stretch 66 hours (LHO-4K)

LLO-4K LHO-4K LHO-2K 3x Coinc.
Duty cycle (cf. S1) 37 (42) 74 (58) 58 (73) 22 (24)
20
S2 Sensitivity and StabilityInspiral Range for
SNR8 with 1.4 - 1.4 M? Inspiral
Virgo Cluster
Preliminary Calibration
Andromeda Galaxy (M31)
Large Magellanic Cloud
Typical 12 hours
21
Third LIGO Science Run (S3)

• October 31, 2003 January 5, 2004
• Three LIGO interferometers, with some
  participation by TAMA and GEO
• Improvements relative to S2
• Sensitivity better by 3-4x for LHO
  interferometers
• Duty cycle improved for LHO interferometers (gt80
  for H1 so far)
• Reduction of acoustic nose coupling (possible
  source of correlated noise at LHO)
• Sensitivity and duty cycle for LLO interferometer
  S2 level

22
Schedule for Full Sensitivity Operation

• TAMA
• Installation of new seismic isolation system in
  2004
• Should lead to improved duty cycle and low f
  sensitivity
• GEO600
• Harder to predict schedule because of new
  technologies
• Observations may take a backseat to technology
  development
• Virgo
• First full interferometer lock within a few
  months
• One year commissioning to bring to full
  sensitivity (my guess!)
• LIGO
• Installation of external preisolator at LLO in
  early 2004
• Full sensitivity operation by the end of 2004

23
Potential for Current Generation of
Interferometers

• My personal assessment
• Binary inspirals
• NS-NS range 20 Mpc
• BH-BH range 100 Mpc
• Continuous waves from neutron stars
• Minimum h few x 10-26
• Stochastic background
• Minimum W 10-6
• Generic bursts
• Minimum EGW lt 1 M? for source at 100 Mpc
• Less certain than other projections